On Thu, 29 Jul 2004 09:12:36 +0300, x wrote:
> "Jan Hidders" <jan.hidders_at_REMOVETHIS.pandora.be> wrote in message
> news:pan.2004.07.28.16.12.50.907297_at_REMOVETHIS.pandora.be...
>>
>> Which problem exactly? To derive the CKs in the projection given a set
>> of CKs in the original relation? That algorithm I already gave in this
>> thread. Or do you want to start from a set of FDs?
>
> From a set of constraints.
> For example FD, MVD, JD, inclusion dependencies, and maybe interval
> constraints, cardinality constraints.
It's always going to be exponential for the simple reason that that is the
potential size of the output. Of the dependencies only the FDs are
actually relevant, you can ignore the rest, and for those you can use the
usual algorithm.
Received on Thu Jul 29 2004 - 13:29:51 CDT
